GGT: Graph-Guided Testing for Adversarial Sample Detection of Deep Neural Network.
Zuohui ChenRenxuan WangJingyang XiangYue YuXin XiaShouling JiQi XuanXiaoniu YangPublished in: CoRR (2021)
Keyphrases
- neural network
- detection method
- detection accuracy
- object detection
- graph structure
- feed forward neural networks
- graph theory
- detection algorithm
- pattern recognition
- structured data
- artificial neural networks
- false positives
- recurrent neural networks
- multi layer
- automatic detection
- graph mining
- graph representation
- random walk
- graph theoretic
- network architecture
- network model
- selecting a small number
- weighted graph
- feed forward
- neural network model
- detection rate
- fault diagnosis
- test cases
- back propagation
- knn
- object recognition
- multi agent